Download pdf - Working Paper No. 881 - Levy Economics Institute · that government finance variables do not appear to exert upward pressure on Indian government bond yields. However, in that study

Working Paper No. 881

The Long-run Determinants of Indian Government Bond Yields

by

Tanweer Akram* Thrivent Financial

Anupam Das

Mount Royal University

January 2017

* Tanweer Akram is director of global public policy and economics at Thrivent Financial. Anupam Das is associate professor in the department of economics, justice, and policy studies at Mount Royal University, Alberta, Canada. Disclaimer: The authors’ institutional affiliations are provided solely for identification purposes. Views expressed are solely those of the authors and the standard disclaimer applies. The views are not necessarily those of Thrivent Financial, Thrivent Investment Management, or any affiliates. This is for information purposes only and should not be construed as an offer to buy or sell any investment product or service. Disclosure: Akram’s employer, Thrivent Financial, invests in a wide range of securities. Asset management services are provided by Thrivent Asset Management, LLC, a wholly owned subsidiary of Thrivent Financial for Lutherans. Note: The data set used in the empirical part of this paper is available upon request to bona fide researchers for replication and verification of the results.

The Levy Economics Institute Working Paper Collection presents research in progress by Levy Institute scholars and conference participants. The purpose of the series is to disseminate ideas to and elicit comments from academics and professionals.

Levy Economics Institute of Bard College, founded in 1986, is a nonprofit, nonpartisan, independently funded research organization devoted to public service. Through scholarship and economic research it generates viable, effective public policy responses to important economic problems that profoundly affect the quality of life in the United States and abroad.

Levy Economics Institute

P.O. Box 5000 Annandale-on-Hudson, NY 12504-5000

http://www.levyinstitute.org

Copyright © Levy Economics Institute 2017 All rights reserved

ISSN 1547-366X

1

ABSTRACT

This paper investigates the long-term determinants of Indian government bonds’ (IGB) nominal

yields. It examines whether John Maynard Keynes’s supposition that short-term interest rates are

the key driver of long-term government bond yields holds over the long-run horizon, after

controlling for various key economic factors such as inflationary pressure and measures of

economic activity. It also appraises whether the government finance variable—the ratio of

government debt to nominal income—has an adverse effect on government bond yields over a

long-run horizon. The models estimated here show that in India, short-term interest rates are the

key driver of long-term government bond yields over the long run. However, the ratio of

government debt and nominal income does not have any discernible adverse effect on yields over

a long-run horizon. These findings will help policymakers in India (and elsewhere) to use

information on the current trend in short-term interest rates, the federal fiscal balance, and other

key macro variables to form their long-term outlook on IGB yields, and to understand the

implications of the government’s fiscal stance on the government bond market.

Keywords: Government Bond Yields; Interest Rates; Monetary Policy; India

JEL Classifications: E43, E50, E60, G10, G12, O16

2

SECTION I: INTRODUCTION

John Maynard Keynes (1930) contends that a central bank’s monetary policy is the most

important driver of long-term interest rates. He believes that the central bank’s actions influence

long-term interest rates, primarily through the effect of the policy rates on short-term interest

rates and other tools of monetary policy. In the General Theory, Keynes (2007 [1936]) reiterates

the importance of central bank’s influence on long-term interest rates, even though he

acknowledges that interest rates have psychological, social, and conventional foundations and

arise from investors’ liquidity preferences.

This paper examines whether Keynes’s conjecture that short-term interest rates are the key driver

of long-term government bond yields holds in India over the long term after controlling for

various key economic factors, such as inflationary pressure and measures of economic activity. It

also appraises if public finance variables, such as the ratio of government debt to nominal

income, have an adverse effect on government bond yields in India. Akram and Das (2015a and

2105b) report that Keynes’s conjectures hold in India for the short-run horizon. They also find

that government finance variables do not appear to exert upward pressure on Indian government

bond yields. However, in that study they do not examine if these results hold over a long horizon.

Understanding the determinants of government bond yields in India over the long-run horizon is

an important question. This question is important not just for scholarly reasons but also for

policy purposes and policy modelling, particularly for discerning the effects of fiscal and

monetary policy on government bond yields. Understanding the drivers of government bond

yields in emerging markets, such as India, is a crucial question for determining the government

finance and macroeconomic policy mix. It is also relevant for fixed income investment and

portfolio allocation, as well as the management of government debt.

This paper investigates whether Keynes’s conjecture that short-term rates drive long-term

interest rates holds over the long-run horizon in the case of India. The paper is organized as

follows. Section II discusses Keynes’s view on interest rates, provides the theoretical framework,

and a simple model. Section III describes the data, the behavioral equations to be estimated, and

3

the econometric methodology applied here. Section IV reports the empirical findings. Section V

analyzes the policy implications of the results and concludes.

SECTION II: THE THEORETICAL FRAMEWORK, A MODEL, AND

INSTITUTIONAL BACKGROUND

The Keynesian Framework

This paper investigates the long-run determinants of Indian government bonds based on

Keynes’s (1930 and 2007 [1936]) view. Keynes holds that the central bank’s actions play the

decisive role in setting the long-term interest rates on government bonds (Kregel 2011). He

argues against the classical view of interest rates based on the loanable funds views, as

represented in Cassel (1903), Marshall (1890), Taussig (1918), and the classical economists.

The central bank’s ability to influence long-term interest rates arises from its ability to set the

policy rates and anchor short-term rates around the policy rates, and to use various other tools of

monetary policy (Keynes 1930). He acknowledges that interest rates have a foundation based on

human psychology, social conventions, herd mentality, and liquidity preferences (Keynes 2007

[1936]). Nevertheless the most immediate and important driver of long-term government bond

yields are the central bank’s actions, as manifested through its ability to: (1) influence the short-

term interest rates by setting the policy rates; and (2) use a wide range of tools of monetary

policy, including expanding and contracting its balance sheet as it deems appropriate. Keynes

relies on Winfried Riefler’s (1930) pioneering empirical analysis of the behavior of interest rates

on U.S. government securities (Kregel 2011). He also observes that current conditions and the

investor’s near-term outlook affect the investor’s long-term outlook. Keynes believes that since

the investor does not have a firm basis for estimating the mathematical expectations of the

unknown and uncertain future, the investor resorts to forming an outlook of the future based on

the past and current conditions. As a result the factors that affect short-term interest rates also

affect long-term interest rates.

4

Keynes’s view on the drivers of long-term government bond yields is in contrast to that of

conventional views in macroeconomics and finance. The conventional view is that government

debts and deficits have a decisive effect on government bond yields. Other things held constant,

if government debt and/or government deficit (both as a share of national income) increases

(decreases), then government bond yields will rise (decline). The conventional view relies on a

loanable funds view of interest rates, whereas for Keynes liquidity preferences and the central

bank’s actions are largely responsible for interest rates as manifested in the yield curve for gilt-

edged (government) securities and other fixed income instruments in an economy.

The conventional view, based on loanable funds theory, is widely represented in the theoretical

and empirical literature. Baldacci and Kumar (2010), Gruber and Kamin (2012), Lam and

Tokuoka (2013), Poghosysan (2014), and Tokuoka (2012) represent the conventional view. In

contrast Akram (2014), Akram and Das (2014a, 2014b, 2015a, and 2015b), and Akram and Li

(2016 and 2017) have argued that short-term interest rates and inflation are the key drivers of

interest rates on government bonds. Moreover, they argue that if other things are held constant,

the government finance variable has hardly any influence on government bond yields. Their view

is based on their interpretation of Keynes and is supported with empirical work on the

determinants of government bond yields in Japan, the United States, and India. As mentioned

earlier, Akram and Das’s (2015a and 2015b) empirical work on India has merely explored the

short-run dynamics. This paper examines whether the same hypothesis holds true for India in the

long run.

A Simple Model of Government Bond Yields

A simple model, based on Akram and Das (2015b) and Akram and Li’s (2017) interpretations of

Keynes’s views, is presented here. To simplify the exposition, a two-period horizon is used.

There are two periods, 1,2. The long-term interest rate on a government bond in period 1 is

; the short-term interest rates on a Treasury bill in period 1 and period 2 are, respectively,

and ; the expected short-term interest rate in period 2 is ; the 1-year, 1-year forward rate is

, ; the term premium is ; the current rate of inflation in period 1 is ; the actual rate of

inflation in period 2 is ; the expected rate of inflation in period 2 is ; the current growth

rate in period 1 is ; the actual growth rate in period 2 is ; the expected growth rate in period

5

2 is ; the government finance variable in period 1 is ; the government finance variable in

period 2 is ; and the expected government finance variable in period 2 is .

The long-term interest rate on the 2-year government bond depends on the short-term interest

rate on Treasury bills in period 1 and the 1-year, 1-year forward rate (equation 1.1). The 1-year,

1-year forward rate is based on the investor’s expectation of the short-term interest rate on

Treasury bills in period 2 and the term premium (equation 1.2). However, the expected short-

term interest rate on Treasury Bills in period 2 and the term premium is a function of the

investor’s expected growth and expected inflation in period 2 (equation 1.3). Hence, the 1-year,

1-year forward rate is merely the sum of the expected short-term interest rate on the Treasury bill

in period 2 and a function of the expected growth rate and expected inflation in the same period

(equation 1.4). This implies that the forward rate is a function of expected short-term interest

rates on the Treasury bill, the expected growth rate, and the expected rate of inflation in period 2

(equation 1.5). Since the long-term interest rate is a function of the short-term interest rate on the

Treasury bill in period 1 and the 1-year, 1-year forward rate (equation 1.6), it follows that the

long-term interest rate is a function of the short-term interest rate in period 1, and a function of

the expected short-term interest rate, the expected growth rate, and the expected rate of inflation

in period 2 (equation 1.7).

Keynes’s view was that the investors resort to the present and the past and rely on their view of

the near-term future to form a view of the long-term future since it is not really possible to have a

proper mathematical expectation of the unknown and uncertain future. Hence, for an investor the

expected short-term interest rates in period 2 are based on the actual short-term interest rates in

period 1 (equation 1.8), the expected growth rate in period 2 is based on the actual growth rate in

period 1 (equation 1.9), and the expected rate of inflation in period 2 is based on the actual rate

of inflation in period 1 (equation 1.10). Similarly the expected government finance variable in

period 2 is based on the government finance variable in period 1 (equation 1.11). These

Keynesian assumptions results in a model (equation 1.12) where the long-term interest rate is a

function of: (1) the current short-term interest rate, the current growth rate, and the current rate of

inflation (equation 1.13); or (2) the current short-term interest rates, the current growth rate, the

current rate of inflation, and the current government finance variable (equation 1.14).

6

The Keynesian view that the investor’s expectation of key economic variables depends largely

on current conditions, or on the investor’s assessment of current conditions, may appear

intriguing and counterintuitive. But if key economic variables follow a Markov process

(equations 1.15, 1.16, 1.17, and 1.18), then the Keynesian view of the trajectory of expected

values of these variables is entirely reasonable. Empirical and behavioral studies of the investor’s

expectations of interest rates and the expectations of the rate of inflation show that Keynes’s

conjectures have considerable support (Clark and Davig 2008; Faust and Wright 2013; and

Mavroedis, Plagborg-Moller, and Stock 2014).

In contrast, under rational expectations, where Lucasian assumptions of perfect foresight hold,

the investor’s expected short-term interest rates, expected growth rate, expected rate of inflation,

and expected government finance variable would equal, respectively, the actual short-term

interest rates, the growth rate, the rate of inflation, and the government finance variable in period

2 (equations 1.19, 1.20, 1.21, and 1.21). This would result in long-term interest rates being a

function of: (1) the current short-term interest rates, the growth rate, and the rate of inflation in

period 2 (equation 1.22); or (2) the current short-term interest rate, the growth rate, the rate of

inflation, and the government finance variable in period 2 (equation 1.23).

The model is represented in the following system of equations:

1 1 1 , equation [1.1]

, equation [1.2]

, equation [1.3]

, , equation [1.4]

, , , equation [1.5]

, , equation [1.6]

, , , equation [1.7]

7

The Keynesian assumptions imply that the following holds:

equation [1.8]

equation [1.9]

equation [1.10]

equation [1.11]

Incorporating Keynesian assumptions into the model leads to the following:

, , , equation [1.12]

, , equation [1.13]

Extending the model to include a government finance variable results in the following:

, , , equation [1.14]

If the variables in period 2 are to follow simple Markov processes, these variables can be

modeled in the following terms:

Λ Λ equation [1.15]




In the above equations, the restrictions on the parameters are as follows: Λ2 1, Λ4 1, Λ6

1, and Λ8 1.

8

It is useful to contrast the Keynesian model with a Lucasian (rational expectations) model. Under

rational expectations:

equation [1.19]

equation [1.20]

equation [1.21]

equation [1.22]

Under a Lucasian assumption, the long-term rates are modeled, respectively, without and with a

government finance variable, as follows:

, , , equation [1.23]

, , , , equation [1.24]

Institutional Background

Akram and Das (2015a and 2015b) provide the institutional background to the monetary policy

framework, government bond market, and monetary-fiscal coordination in India. Yanamandra

(2014) gives additional perspective on monetary policymaking in India in light of economic

reforms, modernization, and recent developments, while Chakraborty (2016) provides a detailed

description and analysis of the country’s fiscal-monetary policy mix and monetary-fiscal

coordination. Jácome et al.’s (2012) survey of global practices of the central bank’s extension of

credit and coordination with the national Treasury includes a description of Indian laws,

regulations, and practices.

India enjoys monetary sovereignty, as defined by Wray (2012). The Government of India issues

its own currency, the rupee. The country’s central bank, the Reserve Bank of India (RBI), sets

policy rates and can use a wide range of monetary policy tools. The RBI enjoys a wide range of

authority and control over the country’s financial system. The Government of India has the legal

and political authority to collect taxes from households, businesses, financial institutions, and

other organizations. The country’s sovereign debt is predominantly issued in its own currency,

9

the rupee. The multifaceted roles played by the RBI in the payment system, monetary policy,

financial stability policy, and policy coordination with the Treasury gives it the operational

ability to influence nominal yields of government bonds by setting and changing short-term

interest rates and using other tools of monetary policy as it deems appropriate.

The next section introduces the behavioral equations, time-series data, and econometric methods

to be used in examining the importance of short-term interest rates, government finance

variables, and other key macroeconomic variables in determining the nominal yields on Indian

government bonds over the long-run horizon.

SECTION III: DATA, BEHAVIORAL EQUATIONS, AND METHODS

Data

For the purpose of econometric estimations, time-series data on the nominal yields of long-term

Indian government bonds (IGBs), short-term interest rates, the rate of inflation, the growth of

industrial production, and government finance variables are used. Nominal yields on Indian

Treasury bills of 3-month maturities are used for the short-term interest rates, while the nominal

yields on IGBs—such as yields on IGBs of 2-year, 3-year, 5-year, 7-year, and 10-year

maturities—are used to represent the long-term government bond yields. Figure 1 shows the

evolution of the nominal yields of IGBs. Figure 2 shows the evolution of short-term interest rates

along with the RBI’s policy rates. The rate of inflation is defined as the year-over-year

percentage change in the Consumer Price Index (CPI) for all items. Growth in industrial

production is the year-over-year change in the index of industrial activity in India. The ratio of

government debt and nominal GDP measures the government variable finance. Data is collected

from Macrobond’s (various years) data services. Table 1 provides a summary of the data and a

detailed description of the variables used here. The monthly time-series dataset runs from

2003M03 to 2015M11, while the quarterly dataset includes time-series variables from 2003Q3 to

2015Q3.

10

Figure 1: The Evolution of Indian Government Bond Yields

Figure 2: Policy Rate and Short-term Interest Rates in India

11

Table 1: Summary of the Data and the Variables Variable

Labels Data Description, Date Range Frequency Sources

Indian Short‐term Interest Rates

TB3M;

TB3M_Q

Government Benchmarks, Auction

Rate, 3 Month,%, Yield;

Jan 1999–Oct 2015;

1Q 1999–3Q 2015

Daily; converted

to monthly;

converted to

quarterly

Reserve Bank of

India; Macrobond

Indian Government Bond Yields

IGB2YR;

IGB2YR_Q

Government bond, 2 year, % yield;

Mar 2003–Oct 2015;

2Q 2003–3Q 2015

Daily; converted

to monthly;

converted to

quarterly

Clearing

Corporation of

India; Macrobond

IGB3YR;

IGB3YR_Q

Government bond,, 3 year, %

yield;

Mar 2003–Oct 2015;

2Q 2003–3Q 2015

Daily; converted

to monthly;

converted to

quarterly

Clearing

Corporation of

India; Macrobond

IGB5YR;

IGB5YR_Q


Mar 2003–Oct 2015;

2Q 2003–3Q 2015

Daily; converted

to monthly;

converted to

quarterly

Clearing

Corporation of

India; Macrobond

IGB7YR;

IGB7YR_Q


Mar 2003–Oct 2015;

2Q 2003–2Q 2015

Daily; converted

to monthly;

converted to

quarterly

Clearing

Corporation of

India; Macrobond

IGB10YR;

IGB10YR_Q

Government bonds, 10 year, %,

yield;

Jan 1999–Oct 2015;

1Q 1999–2 2015

Daily; converted

to monthly;

converted to

quarterly

Clearing

Corporation of

India; Macrobond

Inflation

TCPIYOY;

TCPIYOY_Q;

India, Consumer Price Index, The

Economist, Total, % change, y/y

Jan 2007–Oct 2015;

1Q 2007–2Q 2015

Monthly;

converted to

quarterly

The Economist;

Macrobond

Economic Activity

IPYOY;

IPYOY_Q

Industrial production, % change,

y/y;

Jan 1999–Oct 2015;

1Q 1999–2Q 2015

Monthly;

converted to

quarterly

Central Statistical

Organisation, India;

Macrobond

Government Finance

DEBT_RATIO_

Q

Government debt, % of nominal

GDP;

1Q1999–2Q 2015

Quarterly Indian Ministry of

Commerce &

Industry;

Macrobond

12

Both monthly and quarterly data are used to examine the determinants of the nominal yields of

long-term government bonds. However, it should be noted that Indian government finance data

is available only in the quarterly form. Hence, the-debt-to-GDP ratio is included only in the

quarterly equations.

Behavioral Equations

The following behavioral equations for monthly data and for quarterly data are constructed in

concordance with the model based on the Keynesian framework presented earlier. These

behavioral equations readily lend themselves to empirical testing.

The behavioral equations for the monthly data can be described as follows. The first set of

behavioral equations consists of only one explanatory variable. The long-term interest rate on

government bonds depends only on the short-term interest rate (equation 2.1), inflation (equation

2.2), or the growth rate (equation 2.3). The second set of behavioral equations consists of two

explanatory variables. The long-term interest rate on government bonds depends on the short-

term interest rate and the rate of inflation (equation 2.4) or the short-term interest rates and the

growth rate (equation 2.5). The third set of behavioral equations consists of all three explanatory

variables. The long-term interest rate on government bonds depends on the short-term interest

rate, the rate of inflation, and the growth rate (equation 2.6).

equation [2.1]

equation [2.2]

equation [2.3]

equation [2.4]

equation [2.5]

equation [2.6]

These equations are estimated using monthly data for bond yields of government securities for

various tenors using a variety of indicators of short-term interest rates, the rate of inflation, and

the growth rate.

13

The behavioral equations for the quarterly data incorporate the government finance variables.

The first set of equations consists of just two explanatory variables. The long-term interest rate

depends on: the short-term interest rate and the government finance variables (equation 2.7); the

rate of inflation and the government finance variable (equation 2.8); or the rate of growth and the

government finance variable (equation 2.9). The second set consists of three explanatory

variables: the short-term interest rate, the rate of inflation, and the government finance variable

(equation 2.10); or the short-term interest rate, the growth rate, and the government finance

variable (2.11). The third set consists of four explanatory variables: the short-term interest rates,

the rate of inflation, the growth rate, and the government finance variable (equation 2.12). The

specific-to-general estimation approach for both monthly and quarterly variables is used here.

The following equations are estimated for government bond yields using quarterly data.

equation [2.7]

equation [2.8]

equation [2.9]

equation [2.10]

equation [2.11]

equation [2.12]

These equations are estimated using quarterly data for bond yields of government securities for

various tenors with short-term interest rates, the rate of inflation, the growth rate, and the

government finance variable as explanatory variables.

Econometric Methodology

The first step in the methodology is to examine the nature of the data. The presence of unit roots

in most macroeconomic variables is fairly common (Nelson and Plosser 1982). Hence,

estimating long-run relationships of stationary variables using standard cointegration techniques

(such as Johansen cointegration) is inconsistent. Therefore, unit root tests on the variables used

in this paper are imperative. Both the Augmented Dickey-Fuller (ADF) (Dickey and Fuller 1979

14

and 1981) and Phillips-Perron (PP) (Phillips and Perron 1988) tests are conducted. Different

versions of ADF and PP tests (with no constant and trend, constant and no trend, and constant

and trend) produce similar results. Due to space constraint only the results with no constant and

trend are presented here. (However, the remaining results are available upon request.) The results

for monthly variables are displayed in table 2 and the results for quarterly variables are displayed

in table 3. For the monthly variables, all variables except IPIYOY are non-stationary at levels

and stationary at the first difference. Thus most variables are integrated of order one (I(1)). Both

ADF and PP tests suggest that IPIYOY is stationary at levels, that is, I(0). For quarterly

variables, all variables are found to be non-stationary at levels and stationary at the first

difference, that is, variables are integrated of order one I(1).

Table 2: Unit Root Tests for Monthly Variables Variable Augmented Dickey‐Fuller (ADF) Phillips‐Perron (PP)

IGB2YR 0.07 0.02

ΔIGB2YR ‐11.59*** ‐11.59***

IGB3YR 0.12 0.10

ΔIGB3YR ‐7.59*** ‐11.56***

IGB5YR 0.14 0.12

ΔIGB5YR ‐7.87*** ‐11.43***

IGB7YR 0.14 ‐2.06

ΔIGB7YR ‐7.96*** ‐11.20***

TB3M ‐0.98 ‐0.96

ΔTB3M ‐17.12*** ‐17.15***

TCPIYOY ‐0.57 ‐0.57

ΔTCPIYOY ‐9.55*** ‐9.53***

IPIYOY ‐2.30** ‐7.01***

ΔIPIYOY ‐9.74*** ‐47.61***

Notes: *** and ** indicate statistical significance at the 1 percent and 5 percent level, respectively. Null hypothesis of both ADF and PP tests is that the series contains unit root.

15

Table 3: Unit Root Tests for Quarterly Variables Variable Augmented Dickey‐Fuller (ADF) Phillips‐Perron (PP)

IGB2YR_Q 0.15 0.18

ΔIGB2YR_Q ‐7.48*** ‐7.48***

IGB3YR_Q 0.14 0.36

ΔIGB3YR_Q ‐8.09*** ‐8.25***

IGB5YR_Q 0.11 0.49

ΔIGB5YR_Q ‐8.52*** ‐9.13***

IGB7YR_Q 0.09 0.55

ΔIGB7YR_Q ‐6.76*** ‐9.52***

IGB10YR_Q ‐1.33 ‐1.33

ΔIGB10YR_Q ‐8.82*** ‐8.82***

TB3M_Q ‐0.92 ‐0.96

ΔTB3M_Q ‐8.56*** ‐8.64***

TCPIYOY_Q ‐0.47 ‐0.34

ΔTCPIYOY_Q ‐6.67*** ‐6.76***

DRATIO_Q 0.08 0.18

ΔDRATIO_Q ‐2.61*** ‐11.38***

Notes: ***, **, and * indicate statistical significance at the 1 percent, 5 percent, and 10 percent level, respectively. Null hypothesis of both ADF and PP tests is that the series contains unit root.

To estimate the long-run cointegrating relationships, a number of techniques—including the

autoregressive distributive lag (ARDL) proposed by Pesaran and Shin (1998) and Pesaran, Shin,

and Smith (2001), and the Johansen cointegration proposed by Johansen and Juselius (1990)—

have been used in the existing literature. The ARDL bounds–test approach is based on the

ordinary least square (OLS) estimation of a conditional unrestricted error correction model

(UECM) for cointegration analysis. The ARDL technique is more appealing than the Johansen

cointegration technique because the latter requires that the variables are integrated of the same

order of I(1). The ARDL approach, on the contrary, is not constrained by the outcomes of unit

root tests. It is applicable irrespective of whether the regressors in the model are purely I(0),

purely I(1), or mutually cointegrated. In the present case most variables are I(1) and one is I(0).

Moreover, the ARDL technique allows different variables to take different optimal numbers of

lags, while this is not permitted in the Johansen cointegration approach. Therefore, the ARDL

16

technique, which will accommodate both I(0) and I(1) variables, will be used to estimate the

long-run relationships between long-term government bond yields and other control variables.

SECTION IV: EMPIRICAL RESULTS

The ARDL bounds tests results generated from monthly variables are presented in tables 4–8.

When the short-term interest rate is included with the inflation variable, in most cases the

computed F-statistics based on Wald test exceeds the upper bounds value at the 5 percent level.

In the case of IGB2YR, the computed F-statistics exceeds the upper bounds value at the 10

percent level when the short-term rate is included in the equation with both inflation and an

industrial production index (equation 3.6). The null hypothesis of no cointegration is rejected

whenever the F-statistics value is higher than the upper bounds value. This analysis confirms the

presence of a long-run relationship among long-term government bond yields, short-term interest

rates, the rate of inflation, and the growth in industrial production. It enables estimation of the

long-run coefficients of short-term interest rates and other control variables. The coefficients of

short-term interest rates are always positive and statistically significant at the 1 percent level.

The size of this coefficient tends to be smaller as the tenors of government bonds rise. These

results, thus, suggest that in the long run the short-term interest rates strongly influence the long-

term government bond yields in India.

17

Table 4: ARDL Bounds Test Results for IGB2YR (Monthly Data) Equation F‐statistics

4.1) IGB2YR = β0+β1TB3M 3.93

4.2) IGB2YR = β2+β3TCPIYOY 2.97

4.3) IGB2YR = β4+β5IPIYOY 1.46

4.4) IGB2YR = β6+β7TB3M+β8TCPIYOY 6.52**

4.5) IGB2YR = β9+β10TB3M+β11IPIYOY 2.99

4.6) IGB2YR = β12+β13TB3M+β14TCPIYOY+β15IPIYOY 4.81*

Long‐run Relationships

Variable Equation 4.4 Equation 4.6

TB3M 0.51***

(0.04)

0.51***

(0.05)

TCPIYOY ‐0.01

(0.04)

‐0.00

(0.04)

IPIYOY ‐ ‐0.00

(0.01)

Constant 3.60***

(0.48)

3.60***

(0.54)

Number of Observations 107 105

Notes: ***, **, and * represent 1 percent, 5 percent, and 10 percent levels of significance, respectively. Standard errors are in the parenthesis. Lower bounds values are 6.84, 4.94, and 4.04 for 1 percent, 5 percent, and 10 percent levels of significance, respectively. Upper bounds values are 7.84, 5.73, and 4.78 for 1 percent, 5 percent, and 10 percent levels of significance, respectively.

18


4.7) IGB3YR = β16+β17TB3M 4.60


4.9) IGB3YR = β20+β21IPIYOY 2.03

4.10) IGB3YR = β22+β23TB3M+β24TCPIYOY 8.37***


4.12) IGB3YR = β28+β29TB3M+β30TCPIYOY+β31IPIYOY 6.20**



TB3M 0.39***

(0.04)

0.38***

(0.05)

TCPIYOY ‐0.01

(0.04)

‐0.01

(0.04)

IPIYOY ‐ ‐0.01

(0.01)

Constant 4.74***

(0.47)

4.81***

(0.55)


Notes: *** and ** represent 1 percent and 5 percent levels of significance, respectively. Standard errors are in the parenthesis. Lower bounds values are 6.84, 4.94, and 4.04 for 1 percent, 5 percent, and 10 percent levels of significance, respectively. Upper bounds values are 7.84, 5.73, and 4.78 for 1 percent, 5 percent, and 10 percent levels of significance, respectively.

19


4.13) IGB5YR = β32+β33TB3M 3.84


4.15) IGB5YR = β36+β37IPIYOY 2.37






TB3M 0.26***

(0.04)

0.25***

(0.04)

TCPIYOY ‐0.00

(0.04)

‐0.00

(0.04)

IPIYOY ‐ ‐0.01

(0.01)

Constant 5.86***

(0.43)

5.98***

(0.53)



20


4.19) IGB7YR = β48+β49TB3M 4.02


4.21) IGB7YR = β52+β53IPIYOY 2.59





Variable Equation 4.20 Equation 4.22 Equation 4.24

TB3M ‐ 0.19***

(0.03)

0.18***

(0.04)

TCPIYOY 0.03

(0.08)

0.02

(0.04)

0.01

(0.04)

IPIYOY ‐ ‐ ‐0.01

(0.01)

Constant 7.71***

(0.62)

6.40***

(0.43)

6.53***

(0.52)

Number of Observations 107 107 105


21


4.25) IGB10YR = β64+β65TB3M 4.73

4.26) IGB10YR = β66+β67TCPIYOY 7.51**

4.27) IGB10YR = β68+β69IPIYOY 3.60






TB3M ‐ 0.14***

(0.04)

0.13***

(0.04)

TCPIYOY 0.04

(0.05)

0.03

(0.04)

0.02

(0.04)

IPIYOY ‐ ‐ ‐0.01

(0.01)

Constant 7.74***

(0.45)

6.87***

(0.44)

6.99***

(0.53)


Notes: *** and ** represents 1 percent and 5 percent levels of significance, respectively. Standard errors are in the parenthesis. Lower bounds values are 6.84, 4.94, and 4.04 for 1 percent, 5 percent, and 10 percent levels of significance, respectively. Upper bounds values are 7.84, 5.73, and 4.78 for 1 percent, 5 percent, and 10 percent levels of significance, respectively.

22

Estimated results using quarterly data are presented in tables 9–13. When the short-term 3-month

interest rate is included with inflation and the ratio of public debt to nominal GDP, the computed

F-statistics value is mostly higher than the upper bounds value. Long-run coefficients of the

short-term interest rates are positive when significant. The magnitude of this coefficient lies

between 0.13 to 0.53. The debt-to-nominal-GDP ratio is mostly negative and significant at the 1

percent level, suggesting that in the long run a higher debt ratio tends to reduce the nominal

yields of Indian government bonds, contrary to the conventional wisdom. Quarterly data allows

the use of government finance variables but a clear limitation is that these results are based on a

smaller number of observations.

Table 9: ARDL Bounds Test Results for IGB2YR_Q (Quarterly Data) Equation F‐statistics

4.31) IGB2YR_Q = γ0+ γ1TB3M_Q+ γ2DRATIO_Q 2.67

4.32) IGB2YR_Q = γ3+ γ4TCPIYOY_Q+γ5DRATIO_Q 1.68

4.33) IGB2YR_Q = γ6+ γ7IPIYOY_Q+γ8DRATIO_Q 2.21

4.34) IGB2YR_Q = γ9+ γ10TB3M_Q+ γ11TCPIYOY_Q+γ12DRATIO_Q 1.16

4.35) IGB2YR_Q = γ13+ γ14TB3M_Q+ γ15IPIYOY_Q+γ16DRATIO_Q 2.03

4.36) IGB2YR_Q = γ17+ γ18TB3M_Q+ γ19TCPIYOY_Q+

γ20IPIYOY_Q+γ21DRATIO_Q 1.01

Note: Lower bounds values are 6.84, 4.94, and 4.04 for 1 percent, 5 percent, and 10 percent levels of significance, respectively. Upper bounds values are 7.84, 5.73, and 4.78 for 1 percent, 5 percent, and 10 percent levels of significance, respectively.

23


4.37) IGB3YR_Q = γ22+ γ23TB3M_Q+ γ24DRATIO_Q 5.51**



4.40) IGB3YR_Q = γ31+ γ32TB3M_Q+ γ33TCPIYOY_Q+γ34DRATIO_Q 6.17**






TB3M_Q 0.53***

(0.07)

0.44***

(0.03)

TCPIYOY_Q ‐ 0.00

(0.03)

IPIYOY_Q ‐ ‐

DRATIO_Q ‐2.39***

(0.82)

0.69

(0.61)

Constant 7.36***

(1.55)

3.21***

(0.85)



24





4.46) IGB5YR_Q = γ53+ γ54TB3M_Q+ γ55TCPIYOY_Q+γ56DRATIO_Q 9.00***



γ64IPIYOY+γ65DRATIO_Q 6.63***



TB3M_Q 0.41***

(0.09)

0.26***

(0.04)

0.21***

(0.07)

TCPIYOY_Q ‐ ‐0.03

(0.05)

‐0.11

(0.08)

IPIYOY_Q ‐ ‐ ‐0.03

(0.02)

DRATIO_Q ‐3.06***

(1.04)

1.54

(0.92)

1.67

(1.08)

Constant 9.52***

(1.98)

3.73**

(1.36)

4.67**

(1.83)



25



4.50) IGB7YR_Q = γ69+ γ70TCPIYOY_Q+γ71DRATIO_Q 4.50**

4.51) IGB7YR_Q = γ72+ γ73IPIYOY_Q+γ74DRATIO_Q 4.62**

4.52) IGB7YR_Q = γ75+ γ76TB3M_Q+ γ77TCPIYOY_Q+γ78DRATIO_Q 10.04***





Variable Equation 4.49 Equation 4.50 Equation 4.51 Equation 4.52

TB3M_Q 0.35***

(0.10) ‐ ‐

0.18***

(0.05)

TCPIYOY_Q ‐ 0.02

(0.10) ‐

‐0.04

(0.05)

IPIYOY_Q ‐ ‐ ‐0.02

(0.04) ‐

DRATIO_Q ‐3.22***

(1.14)

1.67

(2.16)

‐4.97***

(1.57)

1.71*

(0.98)

Constant 10.40***

(2.17)

5.18

(3.51)

15.71***

(2.53)

4.27***

(1.43)

Number of

Observations 48 34 48 34


26


4.55) IGB10YR_

Q = γ88+ γ89TB3M_Q+ γ90DRATIO_Q 6.82***

4.56) IGB10YR_Q = γ91+ γ92TCPIYOY_Q+γ93DRATIO_Q 5.51**

4.57) IGB10YR_Q = γ94+ γ95IPIYOY_Q+γ96DRATIO_Q 7.88***

4.58) IGB10YR_Q = γ97+ γ98TB3M_Q+ γ99TCPIYOY+γ100DRATIO_Q 10.66***





Variable Equation 4.55 Equation 4.56 Equation

4.57 Equation 4.58

TB3M_Q 0.29

(0.20) ‐ ‐

0.13**

(0.05)

TCPIYOY_Q ‐ 0.03

(0.08) ‐

‐0.05

(0.06)

IPIYOY_Q ‐ ‐ 0.04

(0.07) ‐

DRATIO_Q ‐5.41***

(2.18)

1.53

(1.78)

‐7.52***

(2.16)

1.75*

(1.02)

Constant 14.67***

(4.42)

5.48*

(2.90)

19.90***

(3.56)

4.85***

(1.48)

Number of

Observations 64 34 64 34


27

These results reinforce the findings in Akram and Das’s (2015a and 2015b) study, which reports

that changes in the short-term interest rates are important determinants of changes in the long-

term government bond yields in India. Whereas Akram and Das (2015a and 2015b) established

the results for the short run, this study extends this for the long run.

SECTION V: POLICY IMPLICATIONS AND CONCLUSIONS

The empirical results support Keynes’s conjecture that a central bank’s actions, through its

influence on short-term interest rates and its use of the tools of monetary policy, are the main

drivers of long-term interest rates. In the case of India, the actions of the RBI affect the long-

term interest rates. The long-term interest rates on IGBs are positively associated with the short-

term interest rates on Indian Treasury bills, after controlling for the relevant variables, such as

the rate of inflation, the growth of industrial production, and the debt ratio. Higher (lower) long-

term interest rates on IGBs are associated with higher (lower) short-term interest rates, higher

(lower) rates of inflation, and a faster (slower) pace of industrial production. The results show

that higher government indebtedness does not have an adverse effect on IGBs’ nominal yields.

These findings concur with the results obtained in Akram and Das’s (2015a and 2015b) studies

of short-term dynamics of IGBs, as well as in Chakraborty (2012) and Vinod, Chakraborty, and

Karun’s (2014) studies, both of which use quite different econometric and statistical methods.

The findings reported in this paper have implications for policy debates in India and other

emerging markets with monetary sovereignty, which issue government debt mostly in its own

currencies. The findings are also relevant for ongoing debates and controversies in fiscal policy,

sustainability of government debt, monetary policy, monetary-fiscal coordination, policy mix

during economic fluctuations, and macroeconomic and monetary theory (Bindseil 2004;

Fullwiler 2008 and 2016; Kregel 2011; Sims 2013a and 2013b; Tcherneva 2011; Woodford

2001; and Wray 2003 and 2012). First, the results show that RBI can exert a strong influence on

IGBs’ yields by affecting the short-term interest rates. The RBI can affect the short-term interest

rates on Indian Treasury bills through setting the repo rate and the reverse repo rate (see figure

2). These findings support Keynes’s conjecture about the influence of a sovereign central bank

28

on long-term interest rates. Second, the results also suggest that, contrary to the conventional

wisdom, higher government indebtedness does not raise nominal yields of IGBs. While this

finding is contrary to the conventional view, which is derived from the loanable funds

perspective, it is in concordance with Keynesian and modern money theory (Fullwiler 2016;

Wray 2003 and 2012), which holds that increased government expenditure results in a rise in the

central bank’s reserves and banking deposits in the financial system because the central bank

credits the banks in the financial system in order to facilitate government borrowing and

expenditures. Third, the results suggest that policymakers in India can use appropriate models—

based on information on the current trend in short-term interest rates, government debt ratios,

and other key macro variables—to form their long-term outlook about IGB yields and

understand the implications of the government’s fiscal stance on the government bond markets.

Of course the use of such models requires judgement and prudence, and carries with it model

risks and limitations.

Keynes claims that the central bank has a decisive influence on the long-term interest rates of

government bonds. He believes that short-term interest rates and other monetary policy actions

drive long-term interest rates and that the investor’s long-run outlook is mostly shaped by the

investor’s near-term outlook and assessment of current conditions. This paper shows that

Keynes’s conjecture has empirical support in India over the long-run horizon. It extends Akram

and Das’s (2015a and 2015b) findings for the short-run horizon to the long-run horizon for the

case of India. It contributes to the nascent literature—such as Akram (2014) and Akram and Das

(2014a and 2014b) on Japan, and Akram and Li (2016 and 2017) on the United States—on this

topic of examining whether Keynes’s conjecture holds in various countries. Further research

should extend this to a wider range of countries, both advanced capitalist economies and

emerging markets and developing areas, and apply a broad spectrum of suitable econometric

methods to establish whether these findings can be generalized and determine under which

institutional contexts these findings are warranted.

29

REFERENCES

Akram, T. 2014. “The Economics of Japan’s Stagnation.” Business Economics 49(3): 156–75.

http://dx.doi.org/10.1057/be.2014.19 Akram, T., and A. Das. 2014a. “Understanding the Low Yields of the Long-Term Japanese

Sovereign Debt.” Journal of Economic Issues 48(2): 331–40. http://dx.doi.org/10.2753/JEI0021-3624480206

———. 2014b. “The Determinants of Long-Term Japanese Government Bonds’ Low Nominal

Yields.” Levy Economics Institute Working Paper No. 818. Annandale-on-Hudson, NY: Levy Economics Institute of Bard College.

———. 2015a. “Does Keynesian Theory Explain Indian Government Bond Yields?” Levy

Economics Institute, Working Paper No. 834. Annandale-on-Hudson, NY: Levy Economics Institute of Bard College.

———. 2015b. “A Keynesian Explanation of Indian Government Bond Yields.” Journal of Post

Keynesian Economics 38(4): 565–87. http://dx.doi.org/10.1080/01603477.2015.1090294 (accessed November 25, 2016).

Akram, T., and H. Li. 2016. “The Empirics of Long-Term U.S. Interest Rates.” Levy Economics

Institute, Working Paper No. 863. Annandale-on-Hudson, NY: Levy Economics Institute of Bard College.

———. 2017. “What Keeps Long-Term U.S. Interest Rates So Low?” Economic Modelling

60(January): 380–90. http://dx.doi.org/10.1016/j.econmod.2016.09.017 Baldacci, E., and M. Kumar. 2010. “Fiscal Deficits, Public Debt, and Sovereign Bond Yields.”

IMF Working Paper No. 10/184. Washington, DC: International Monetary Fund. http://www.imf.org/external/pubs/ft/wp/2010/wp10184.pdf

Bindseil, U. 2004. Monetary Policy Implementation: Theory, Past, and Present. Oxford and New

York: Oxford University Press. Cassel, G. 1903. Nature and Necessity of Interest. London and New York: The Macmillan

Company. Chakraborty, L.S. 2012. “Empirical Evidence on Fiscal Deficit—Interest Rate Linkages and

Financial Crowding Out.” Levy Economics Institute Working Paper No. 744. Annandale-on-Hudson, NY: Levy Economics Institute of Bard College.

———. 2016. Fiscal Consolidation, Budget Deficits and the Macro Economy. Thousand Oaks,

CA: Sage Publications Pvt. Ltd.

30

Clark, T.E., and T. Davig. 2008. “An Empirical Assessment of the Relationships among Inflation and Short- and Long-Term Expectations.” Research Working Paper RWP 08-05 (November). Kansas City: Federal Reserve Bank of Kansas City. https://www.kansascityfed.org/publicat/reswkpap/pdf/rwp08-05.pdf

Dickey, D.A., and W.A. Fuller. 1979. “Distribution of the Estimators for Autoregressive Time

Series with a Unit Root.” Journal of the American Statistical Association 74(366): 427–31. http://dx.doi.org/10.1080/01621459.1979.10482531

———. 1981. “Likelihood Ratio Statistics for Autoregressive Time Series with a Unit Root.”

Econometrica 49(4): 1057–72. http://www.jstor.org/stable/1912517 Faust, J., and J.H. Wright. 2013. “Forecasting Inflation.” In Graham Elliot and Allan

Timmermann (eds.), Handbook of Economic Forecasting, vol. 2A. Amsterdam: Elsevier. Fullwiler, S. 2008. “Modern Central Bank Operations: The General Principles.” Unpublished

manuscript. Available at: http://ssrn.com/abstract=1658232 (accessed November 25, 2016).

———. 2016. “The Debt Ratio and Sustainable Macroeconomic Policy.” World Economic

Review 7(July): 12–42. Gruber, J.W., and S.B. Kamin. 2012. “Fiscal Positions and Government Bond Yields in OECD

Countries.” Journal of Money, Credit, and Banking 44(8): 1563–87. http://dx.doi.org/10.1111/j.1538-4616.2012.00544.x

Hansen, L.P. 1982. “Large Sample Properties of Generalized Method of Moments Estimators.”

Econometrica 50(4): 1029–54. http://www.jstor.org/stable/1912775 Jácome, L.I., M. Matamoros-Indorf., M. Sharma., and S. Townsend. 2012. “Central Bank Credit

to the Government: What Can We Learn from International Practices?” IMF Working Paper No. 12/16. Washington, DC: International Monetary Fund. https://www.imf.org/external/pubs/ft/wp/2012/wp1216.pdf

Johansen, S., and K. Juselius. 1990. “Maximum Likelihood Estimation and Inference on

Cointegration—with Applications to the Demand for Money.” Oxford Bulletin of Economics and Statistics 52(2): 169–210. http://dx.doi.org/10.1111/j.1468-0084.1990.mp52002003.x

Keynes, J.M. 1930. A Treatise on Money, Vol. II: The Applied Theory of Money. London:

Macmillan. ———. 2007 [1936]. The General Theory of Employment, Interest, and Money. New York:

Palgrave Macmillan.

31

Kregel, J. 2011. “Was Keynes’ Monetary Policy À Outrance in the Treatise, A Forerunner of ZIRP and QE? Did He Change his Mind in the General Theory?” Levy Economics Institute, Policy Note No. 4. Annandale-on-Hudson, NY: Levy Economics Institute of Bard College.

Lam, R.W., and K. Tokuoka. 2013. “Assessing the Risks to the Japanese Government Bond

(JGB) Market.” Journal of International Commercial and Economic Policy 4(1): 1350002–1-1350002-15. http://dx.doi.org/10.1142/S1793993313500026

Marshall, A., 1890. Principles of Economics. London and New York: Macmillan and Company. Macrobond. various years. Macrobond subscription services. (accessed May 31, 2016). Mavroedis, S., M. Plagborg-Moller, and J.H. Stock. 2014. “Empirical Evidence on Inflation

Expectations in the New Keynesian Phillips Curve.” Journal of Economic Literature 52(1): 124–88. http://dx.doi.org/10.1257/jel.52.1.124

Nelson, C.R., and C.R. Plosser. 1982. “Trends and Random Walks in Macroeconmic Time

Series: Some Evidence and Implications.” Journal of Monetary Economics 10(2): 139–62. http://dx.doi.org/10.1016/0304-3932(82)90012-5

Pesaran, M.H., and Y. Shin. 1998. “An Autoregressive Distributed-Lag Modelling Approach to

Cointegration Analysis.” Econometric Society Monographs 31: 371–413. Pesaran, M.H., Y. Shin, and R.J. Smith. 2001. “Bounds Testing Approaches to the Analysis of

Level Relationships.” Journal of Applied Econometrics 16(3): 289–326. http://dx.doi.org/10.1002/jae.616

Phillips, P.C.B., and P. Perron. 1988. “Testing for a Unit Root in Time Series Regression.”

Biometrika 75(2): 335–46. http://dx.doi.org/10.1093/biomet/75.2.335 Poghosyan, T. 2014. “Long-Run and Short-Run Determinants of Sovereign Bond Yields in

Advanced Economies.” Economic System 38(1): 100–14. http://dx.doi.org/10.1016/j.ecosys.2013.07.008

Riefler, W.W. 1930. Money Rates and Money Markets in the United States. New York and

London: Harper & Brothers. Sims, C. 2013a. “Paper Money.” American Economic Association Presidential Lecture, January

14. http://sims.princeton.edu/yftp/PaperMoney/PaperMoneySlides.pdf (accessed November 25, 2016).

———. 2013b. “Paper Money.” American Economic Review 103(2): 563–84.

http://dx.doi.org/10.1257/aer.103.2.563

32

Taussig, F.W. 1918. Principles of Economics, volume 2 (second revised edition). New York: The Macmillan Company.

Tcherneva, P.R. 2011. “Bernanke’s Paradox: Can He Reconcile His Position on the Federal

Budget with His Recent Charge to Prevent Deflation?” Journal of Post Keynesian Economics 33(3): 411–34. http://dx.doi.org/10.2753/PKE0160-3477330301

Tokuoka, K. 2012. “Intergenerational Implications of Fiscal Consolidation in Japan.” IMF

Working Paper, No. 12/197. Washington, DC: International Monetary Fund. http://www.imf.org/external/pubs/ft/wp/2012/wp12197.pdf

Vinod, H.D., L. Chakraborty, and H. Karun. 2014. “If Deficits Are Not the Culprit, What

Determines Indian Interest Rates? An Evaluation Using the Maximum Entropy Bootstrap Method.” Levy Economics Institute Working Paper No. 811. Annandale-on-Hudson, NY: Levy Economics Institute of Bard College.

Woodford, M. 2001. “Fiscal Requirements for Price Stability.” Journal of Money, Credit and

Banking 33(3): 669–728. http://www.jstor.org/stable/2673890 Wray, L.R. 2003. Understanding Modern Money: The Key to Full Employment and Price

Stability. Paperback edition. Cheltenham, UK and Northampton, MA: Edward Elgar. ———. 2012. Modern Money Theory: A Primer on Macroeconomics for Sovereign Monetary

Systems. New York: Palgrave Macmillan. Yamanadra, S. 2014. “Minsky, Monetary Policy and Mint Street: Challenges for the Art of

Monetary Policymaking in Emerging Economies.” Levy Economics Institute, Working Paper No. 820. Annandale-on-Hudson, NY: Levy Economics Institute of Bard College.